HN user
Very interesting video!
In the slowed version it seemed like the operations were fully sequential, I think they might be able to achieve a shorter time by overlapping some operations and potentially with edge-cutting too
In the slow-montion footage of the shared faster Mitsubishi robot you can see it's doing some operations in parallel (but not edge-cutting)
Valid point, but then I wouldn't consider classical mathematics to be a science
The profusion of jaded nerds, although saddening at times, seems to be pushing science forward. I have a feeling that a prolonged sense of "Awe" can hinder progression at times, and the lack of it is usually a sign of the adaptability of a group (how quick new developments are normalized?)
reminds me of the rabbit hole sessions I used to fall into in https://wiki.c2.com/ (a merge of the two interfaces (chat and context rabbit hole window thing) would be perfect for me)
Did you experience any performance degradation when using alacritty + tmux vs alacritty alone?
Surprisingly in some tests [1], foot was rendering faster than alacritty for me (Wayland, Nvidia GPU)
https://github.com/alacritty/alacritty/issues/289#issuecomme...
I'm wondering the same, but also wonder if the situation off the coast of Yemen and Iran's recent response to Israel bombing their embassy made the conflict partially international?
He meant something more meta I believe. Knowing you are a monkey is one thing, and knowing that you know you are a monkey is a another thing. It's about being cognisant of the fact that there is something called knowledge and you have it
That's really cool, do you think this might be the basis for potential natural language navigation? (when going over a document, instead of having to search by keyword or regex, one can search for more complicated concepts using English)
If not, what extra work is needed to bring it to that level?
Side related question: are there content licenses coming up that are similar in spirit to what the GPL is but targeted at AI training? (E.g. if this piece of content was used in training an AI that was to be used commercially, the AI's weights must be published)
That was my point. Whereas if you are using a service from IBM as an enterprise, you would be able to sue them
Can I sue Lamma and Mistral if things go wrong?
That's the thing, that's not a common denominator if you really know what each group stands for and promotes. There are christians in Gaza under Hamas. There are different Muslim denominations in Gaza under Hamas. The other groups you are mentioning wouldn't allow this, they kill other Muslims for being different, and they attack christians. There is a huge contrast. What you call a common denominator is false. There are many "unbelievers" living fine in Gaza and in Palestine in general (ignoring that they are also being victims of Israeli attacks)
It's even banned in some Arab countries for being a security risk.
If I recall Egypt partially did that after the military coup, and some Gulf countries. None of which are known for being kind to criticism. So this point is more validating of Aljazeera's position as relatively honest journalism in the a region full of dictatorships and corrupt governments.
Their brethren? I believe you are uneducated if you are referring to the group I assume you are. Especially given the politics of the region and the differences/animosity between those groups
Neat Idea!
I was wondering, are there any security guarantees for the providers? Assume I have a small GPU cluster at home, if I rent out my GPUs, what sort of access should I accept the renters to have? Only GPU kernels would be sent to machine? Or will they have a limited permission user access on my cluster?
Also instead of having the operatos openning ports in their routers, were there any considers of adding them to a private network for a more seamless experience? (Nebula/headscale and the likes)
While true, i think this answer misses the hypocrisy of the situation.
1. The same people who were supportive of the previous protests, are today's people who suggest "don't do it at work". 2. Regardless of whether Israel is commiting war crimes or not, if one is to be consistent, they must advocate that both the China censorship and genocide protests be done outside of work. 3. The main takeaway here is the bias (justified or not), and disappoint and disturbing hypocrisy in the reaction to the aforementioned protest
But even without going this far (with integrating various other specialized or having an LLM use them when required), an LLM is probably able to recognize a sudoku puzzle when it sees one, and even tho it itself can't solve it, I think it can easily write the code that would solve sudoku. So instead of hooking it to a set of pre built models, it might be enough to hook it to a python interpreter
Of course! It's from a book containing a collection of old tales. It's called "one thousand tales and a tale" (the book only contains a hundred)
You can find it here. The tale I referenced is on page 61.
(Unfortunately I didn't find an English copy of the book but there might be one)
Reminds me of an old Arab tale (paraphrased, from memory, translated):
Once upon a time a man wanted to be handed the title of a poet, so he went to the king of poets at his time, and asked him to give him the title of a poet. The king of poets asked the man, "Do you memorize tens of thousands of the best poetry?", the man said "I do not", the king of poems asked of him: "go memorize them then come back to me". The man went to a faraway isolated monastary, spend a couple of years memorizing the poems, then came back to the king of poets shouting in excitement "I've memorized the tens of thousands of poems! Can you hand me the title of a poet now?". The king of poets asked of him "Now forget them".
Only after years again spent in the monastery forgetting the poems, was the man given the title of a poet.
(Side note: this story also kinda remind me of LLMs, and how they kind of "memorize" text initially but the more data is poured in the more they "forget" the exact texts)
Technically there are cases in which you can. The repeated values will act like separators that you cannot act on (because once you act on one, as you said one cannot create the final list again), but then it becomes multiple instances of the puzzle which seems pointless
Wonderful work!
is it possible to make it only use a subset of the web? (Only sites that I trust and think are relevant to producing an accurate answer), and are there ways to possibly make it work offline on pre installed websites? (wikipedia, some other wikis and possibly news sites that are archived locally), and how about other forms of documents? (books and research papers as pdfs)
I see, but I was wondering about the partial transferability of this feature to other LLMs
But fair enough, context length is key in this scenario
How would the results compare to:
1. Video frames are sampled (based on frame clarity)
2. The images are fed to OCR, with their content outputed as:
Frame X: <content of the frame>
3. The accomulated text is given to an average LLM (Mistral) and asked the same request mentioned by the author (creating a JSON file containing book information)
Wouldn't we get something similar? maybe if a more sophisticed AI is used? So the monopoly on Gemini Pro for video processing (specifically when it comes to handling text present inside the video) is not really a sustainable advantage? or am I missing something (as this is something beyond just a fancy OCR hooked into a LLM? as the model would be able to tell that this text is on a book for instance?)
Can someone explain the meaning of this to me?
Is the encounter with the vampire similar to forgetting why one came into a room because when encountering a vampire in a newly entered room you forget everything before that and focus on the vampire?
Also for none paid subscribers, is this the full article?
(Does it end with:
"After the break we’ll discuss GPU instruction sets and registers, memory and software." (also side question, what break?) )?
Maybe people reacted this way because they assumed this was the full article
I wholeheartedly agree. For those interested check out the hard problem of consciousness
And I have a thought I would like to share here, an attempt at arguing for why it might be impossible for a physical pure materialistic model of the world to explain consciousness
In a pure physical materialistic world, we can represent everything as laws, interactions between materials, guided by a set of rules. That's it.
If we had a absolute knowledge, we can represent those rules, all of them, as a set of mathematical and logical equations.
I reduce the material world to a mathematical world of rules and laws, an abstract world.
Saying consciousness could arrise from a pure material world, is akin to saying the mathematical equations and laws, combined, can be conscious. As in, there is a system of rules that if tuned in a certain way can think. Since rules are abstract, and by definition rules don't think. No combination of them no matter how complicated can think
The initial rebuttle I have in my mind is that "well you are conflicting the system with the objects governed by it", I feel there is a better way to phrase the initial proposal to address this issue, but I just wanted to throw this idea out there
Such works might be quite useful for natural language AI infused code editing as there would be a coherent interface for the AI to give commands to and a smaller room of error (given that the action space is smaller than when a standard text editor is used)
Greetings
Unfortunately my probability skills are quite bad and i easily get confused, could someone explain where my reasoning fails?
I assumed since we have a uniform distribution of red to green balls, then the expected value would be 50 red balls.
So I assumed I can use this expected distribution to figure out the probability the question asked
in an urn where there is 50 red balls and 50 green balls, and we draw one red ball. The ball we draw next is more likely to be green (since we have 49 red balls vs 50 green balls)
Is my mistake that I took the expected value of the uniform distribution before the first ball was drawn? And I should think of the expected distribution of the urn given the first ball drawn was red? If so how would I go about that probabilistically? (i.e. how would I calculate
Exp [ n given first ball is red ] )
I can see that if the first ball was red, this removes one of the possible values (all balls are green)
And I can see how from all possible urns, the more they have red balls, the higher the likelihood the first ball drawn is red. I would assume this will skew the expected value somehow.
Is it like
``` expVal = 0
for all possible n (101 possible values): # possibility of first ball being red when we have n red balls * n * probability of n?
expVal += (n / 100) \* n \* (1/101)
I'm sure the above is wrong, because this will just decrease the expected value where I am expecting an increase but I'm not sure how to model it
I would appreciate any pointers/help
I still this still ignores important points, and ill make some comments and questions, some of which might be pedantic so feel free to ignore them
1. Regardless of the method, it was fairly accurate for years, and matched closely to Israeli estimates of casualties in prior conflicts.
2. The US, Israeli's largest ally, is not doubting those numbers. This alone says alot given that the US from a strategic perspective would want to present Israel in the best possible light they can given the majority of the globe is heading towards leaving Israel in isolation.
3. The Israeli claim against Almamadani hospital incident is placing the blame on Islamic Jihad (not Hamas) (pedantic point)
4. The issue of Almamadani hospital incident is still not settled, especially as the Israeli claimes have been debunked by multiple entities most notably the New York times. [1]
5. Ignoring the questionable numbers of this specific incident, because you might be right, but its interesting that the US's admission that the death toll coming from Gaza is accurate came at a date way after the Al mamadani hospital incident (my claim would be that this incident has been taken into consideration by the US officials when they admitted the accuracy of the death toll coming from Gaza). This paired with the US's strategic support of Israel makes their admission that the death toll is accurate is way more trust worthy than possibly exaggerated COVID death tolls, as in one case its an admission playing against the admitter, whereas the other case the admission is in favor of the admitter.
But thank you for sharing the Twitter thread, I'll investigate it and look into other sources as well
[1] https://www.nytimes.com/2023/10/24/world/middleeast/gaza-hos...